University of Groningen Rational clinical examination of the critically ill patient Hiemstra, Bart

(1)

Rational clinical examination of the critically ill patient

Hiemstra, Bart

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Hiemstra, B. (2019). Rational clinical examination of the critically ill patient. Rijksuniversiteit Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

531658-L-bw-Hiemstra 531658-L-bw-Hiemstra 531658-L-bw-Hiemstra 531658-L-bw-Hiemstra Processed on: 5-6-2019 Processed on: 5-6-2019 Processed on: 5-6-2019

Processed on: 5-6-2019 PDF page: 145PDF page: 145PDF page: 145PDF page: 145

145

6

Critical Care Medicine 2019; in press

Clinical examination for the prediction

of mortality in the critically ill:

the Simple Intensive Care Studies-I

Hiemstra B, Eck RJ, Wiersema R, Kaufmann T, Koster G, Scheeren TWL, Snieder H, Perner A, Petillä V, Wetterslev J, Keus F, van der Horst ICC, SICS Study Group

(3)

146

Abstract

Objectives

Caregivers use clinical examination to timely recognise deterioration of a patient, yet data on the prognostic value of clinical examination are inconsistent. In the Simple Intensive Care Studies-I (SICS-I), we evaluated the association of clinical examination findings with 90-day mortality in critically ill patients.

Design

Prospective single centre cohort study. Setting

Intensive care unit (ICU) of a single tertiary care level hospital between 27 March 2015 and 22 July 2017.

Patients

All consecutive adults acutely admitted to the ICU and expected to stay for at least 24 hours. Interventions

A protocolised clinical examination of 19 clinical signs conducted within 24 hours of admission. Measurements

Independent predictors of 90-day mortality were identified using multivariable logistic regression analyses. Model performance was compared to established prognostic risk scores using area under the receiver-operating-curves (AUROC). Robustness of our findings were tested by internal bootstrap validation and adjustment of the threshold for statistical significance.

Main results

A total 1,075 patients were included, of whom 298 (28%) had died at 90-day follow-up. Multivariable analyses adjusted for age and noradrenaline infusion rate demonstrated that the combination of higher respiratory rate, higher systolic blood pressure, lower central temperature, altered consciousness, and decreased urine output were independently associated with 90-day mortality (AUROC 0.74; 95% CI 0.71-0.78). Clinical examination had a similar discriminative value as compared to the SAPS-II (AUROC 0.76; 95% CI 0.73-0.79; p=0.29) and APACHE-IV (AUROC 0.77; 95% CI 0.74-0.80; p=0.16) and was significantly better than the SOFA (AUROC 0.67; 95% CI 0.64-0.71; p<0.001).

Conclusions

Clinical examination has reasonable discriminative value for assessing 90-day mortality in acutely admitted ICU patients. In our study population a single, protocolised clinical examination had similar prognostic abilities compared to the SAPS-II and APACHE-IV and outperformed the SOFA score.

(4)

147 Clinical examination for predicting mortality

Introduction

Patients acutely admitted to the Intensive Care unit (ICU) have a high mortality and survivors may suffer from long-term morbidity and reduced quality of life.1,2_{These critically ill patients frequently} present with clinical signs of circulatory shock such as low blood pressure, oliguria, and skin mottling. These signs often guide treatment, assuming they indicate vital organ hypoperfusion and are associated with increased mortality.3-7_{Indeed, guidelines on the management of shock} recommend treating patients based on clinical examination, supplemented with critical care ultrasonography (CCUS).8

Data on the prognostic value of clinical examination findings are inconsistent. Previous studies have identified different predictors of mortality such as low blood pressure,6,9,10_oliguria,3,5 prolonged capillary refill time,11,12_{and skin mottling.}4,13_{They often evaluated one or two clinical} signs in isolation, instead of assessing a combination of signs and symptoms, which would more accurately reflect daily clinical practice. Furthermore, most studies had relatively small sample sizes or included a selected subgroup such as patients with sepsis, cardiogenic shock, or severe trauma (Table S1).

The prognostic value of clinical examination remains to be established in a large, consecutive cohort of critically ill patients. Compared to well-established prognostic scores which are complex to calculate and unsuited for individual patient prognostication,14,15_{a simple bedside} clinical examination might better inform caregivers in their decision making. Accordingly, our aim was to evaluate which clinical examination findings were independently associated with 90-day mortality in acutely admitted ICU patients. In addition, we hypothesised that combined clinical examination findings would have similar prognostic value compared to existing prognostic scores.

Methods

Design, setting and patients

The prospective, observational, single-centre Simple Intensive Care Studies-I (SICS-I) was conducted following a prewritten protocol and statistical analysis plan (SAP; see Page 67 or clinicaltrials.gov: NCT02912624). All consecutive patients admitted to the intensive care unit (ICU) of the University Medical Center Groningen (UMCG) were eligible for inclusion. Adult patients who had an unplanned ICU admission and were expected to stay for at least 24 hours were included. Patients were excluded if their ICU admission was planned pre-operatively, if acquiring research data interfered with clinical care due to continuous resuscitation efforts (e.g. mechanical circulatory support), or if informed consent was not provided. In unresponsive patients informed consent was first obtained from the legal representatives and at a later time if the patient recovered consciousness. If the patient died before consent was obtained, the study data was used and legal representatives were informed on the study. The local institutional review board approved the study (M15.168207).

(5)

148

All included patients underwent clinical examination followed by CCUS within the first 24 hours of their ICU admission (Table S2). Researchers conducted the clinical and CCUS examinations and their findings were not revealed to caregivers.

Clinical examination

All clinical examinations were standardised and cut-off values for abnormal clinical signs were predefined in the protocol (clinicaltrials.gov; NCT02912624). A total of 19 clinical signs per patient were recorded (Table S2). Respiratory rate, heart rate and rhythm, arterial blood pressures and central venous pressures were recorded from the bedside monitor. Patients were auscultated for the presence of cardiac murmurs and crepitations. Clinical signs reflecting organ perfusion were obtained from the three organs readily accessible to clinical examination: cerebral (mental status), renal (urine output) and skin perfusion (capillary refill time (CRT), central-to-peripheral temperature difference (ΔTc-p) and skin mottling). Mental status was assessed according to the categories ‘Alert’, ‘responsive to Voice’, ‘responsive to Pain’ and ‘Unresponsive’ (AVPU-scale) and was scored irrespective of sedation use. Urine output was scored one and six hours prior to the clinical examination, adjusted for body weight, and considered decreased if < 0.5 ml∙kg-1_∙h-1_{. CRT was} the time for skin colour to fully return after applying firm pressure at the sternum, index finger, and knee for 15 seconds and considered prolonged if >4.5 seconds.16_{ΔTc-p was the difference} between central temperature measured by a bladder thermistor catheter and peripheral temperature measured by a skin probe on the big toe and dorsum of the foot and considered abnormal if >7 °C.17,18_{The degree of skin mottling was rated at the knee according to a score from} 0 to 5, where 0-1 was regarded as mild, 2-3 as moderate and 4-5 as severe mottling.19

Outcome definition

The primary outcome (dependent) variable was 90-day all-cause mortality obtained through the municipal record database. Sensitivity analyses were conducted using all-cause mortality at 7- and 30-day follow-up.

Sample size and missing data

The sample size was based on the estimation that half of the number of acute ICU admissions per year (n = 1,500) would fulfil the inclusion criteria. The potentially detectable difference was calculated using skin mottling as an example for the case inclusion exceeded 1,000 patients: a significant mortality difference of 9% for skin mottling with 84% power and a maximal type 1 error risk of 0.015 could be detected.20_{Missing values were considered missing at random because} these depended on other observed patient characteristics (such as age, mechanical ventilation) and a significant Little’s test.21_{Multiple imputations (20 times) for missing data were conducted} and parameter estimates and standard errors were combined using Rubin’s formula.22,23

(6)

149 137 Excluded, reason:

 80 No CCUS possible  40 Logistic reasons  17 Other reasons

1075 Included in SICS-I cohort 1212 Fulfilled inclusion criteria

5587 Patients admitted to ICU

 124 Unable to provide informed consent - 82 Refused to participate - 26 Not mentally competent - 16 Serious language barrier  54 Died prior to inclusion  40 Age < 18 years

 12 Continuous resuscitation efforts

1442 Assessed for eligibility

 2977 Elective admissions  1168 Discharged within 24 hours

Figure 1. Flow diagram of the Simple Intensive Care Studies-I (SICS-I) cohort. Abbreviations: ICU, intensive care CCUS, critical care ultrasonography

Analytical approach

The aims of our primary analyses were twofold: first, a multivariable logistic regression analysis was conducted to identify the clinical examination findings that independently predict mortality at 90- day follow-up. Second, the discriminative performance of this model was compared to that of the simplified acute physiology score (SAPS-II), acute physiology and chronic health evaluation (APACHE-IV) and sequential organ failure assessment (SOFA). Analyses were conducted with Stata version 15.1 (StataCorp, College Station, Texas, USA) on the imputed dataset following our published SAP (see Page 67).

(7)

150

Model development and validation

Unadjusted and age- and sex adjusted regression analyses were conducted on 19 clinical signs. A p < 0.25 threshold was used for inclusion in the multivariable models, which was constructed using forward stepwise regression by adding blocks of variables. The multivariable model was adjusted for age (covariate) and norepinephrine infusion rate (mediator) under the pathophysiological mechanism that norepinephrine alters most clinical signs. The final model was internally validated with bootstrap sampling. For bootstrap sampling, 1,075 cases were repeatedly drawn with replacement from the imputed dataset (see Figure S1 for a more in-depth explanation). In total, one-hundred bootstrap samples were drawn and the final model was re-constructed in each sample. Each variable from the final model was considered internally validated if it was significant in at least 80 of the 100 bootstrapped models.20,24_{Calibration of the multivariable models was} checked with calibration plots and Hosmer-Lemeshow tests. Discrimination of the final model was evaluated with receiver operating characteristic (ROC)-curves.25_{Dominance analysis was used to} determine the relative importance of independent variables in each multivariable model.26_Our multivariable model was compared to the SAPS-II (reference model), APACHE-IV and SOFA scores by 1) analysing differences between the area under the ROC curves (AUC) using the method proposed by Delong et al.,27_{and by 2) constructing reclassification tables and calculating the net} reclassification improvement.28

Sensitivity and subgroup analyses

In sensitivity analyses we assessed whether the statistically significant predictors of 90-day mortality were also predictive of 7 and 30-day mortality. Time-dependency was also investigated by conducting a multivariable Cox regression analysis on 90-day mortality.

Two planned subgroup analyses were conducted, in which only the clinical examination findings that were statistically significant in the primary analysis were evaluated. First, patients were stratified by vasopressor use. Second, patients were stratified by underlying pathology that could influence the clinical measurements, i.e. acute liver failure or post orthotopic liver transplantation (OLT), heart failure, septic shock, cardiac arrest, and central nervous system (CNS) pathology.

Statistical significance

The SICS-I was designed to address multiple hypotheses on six different outcomes and, therefore, the mortality outcome was adjusted for multiple hypothesis testing.29_{Supplements 1 contain} the details or our SAP, but in short a p-value of 0.015 indicated statistical significance and p-values between 0.015 and 0.05 indicated suggestive significance with an increased family-wise error rate.20,30_{For our secondary (subgroup) analyses, a p-value below 0.05 indicated statistical} significance due to the hypothesis-generating purpose. Accordingly, primary analyses are presented with 98.5% CIs and secondary (subgroup) analyses with 95% CIs.

(8)

151

Table 1. Clinical characteristics

Variable All patients

N = 1075 Age [years] 62 ± 15 Sex, male 674 (63%) Admission type Medical 713 (66%) Acute surgery 362 (33%)

Admission diagnosis by organ system

Cardiovascular 318 (30%) Gastrointestinal 167 (16%) Genito-urinary 23 (2%) Hematological 19 (2%) Metabolic 22 (2%) Musculoskeletal/skin 13 (1%) Neurological 143 (13%) Respiratory 229 (21%) Transplant 58 (5%) Trauma 82 (8%) Subgroups

Acute heart failure 63 (11%)

Cardiac arrest 125 (21%) CNS pathology 144 (24%) Liver failure 54 (9%) Sepsis 206 (35%) Prognostic scores APACHE-IV 76 ± 29 SAPS-II 46 ± 17 SOFA 8 ± 5

Abbreviations: PEEP, positive-end-expiratory pressure; CNS, central nervous system; APACHE, acute physiology and chronic health evaluation; SAPS, simplified severe acute physiology score; SOFA; sequential organ failure assessment.

Amendments to the statistical analysis plan

For our primary analysis multivariable logistic regression analyses was used instead of Cox regression because the outcome (90- day mortality) was fixed, time-to-event was considered less relevant, and our statistical methods would be more in line with that in the literature. Findings of the Cox regression analyses are reported in the supplements.

We intended to conduct a multivariable regression analysis of clinical examination findings adjusted for the SAPS-II. Since the SAPS-II also contains various clinical examination findings, we realised that such a model would have little clinical relevance and instead used this score as the reference model. We compared the performance of our clinical examination model to the SAPS-II, APACHE-IV, and SOFA.

Table 1. Clinical characteristics

Abbreviations: expiratory pressure; CNS, central nervous system; APACHE, acute physiology and chronic health evaluation; SAPS, simplified severe acute physiology score; SOFA; sequential organ failure assessment.

(9)

152

Results

Patient characteristics and outcome

A total of 1,442 patients were assessed for eligibility between 27 March 2015 and 22 July 2017. The inclusion criteria were fulfilled in 1,212 patients of whom 137 were not included for various reasons (Figure 1). In the final analysis 1,075 patients (89%) were included. The median time from ICU admission to inclusion was 15 hours (IQR 8 – 20 hours). The proportion of missing values per variable (Table S2) and per case (Figure S2) is presented in the supplements. One-third of the patients were admitted after acute or complicated surgery and the most common admission diagnoses were of cardiovascular or respiratory origin (Table 1).

After 90 days 298 patients (28%) had died, and 8 patients (1%) were lost from follow-up due to emigration to or residence in another country. Patients who died within 90 days were significantly older, were more often mechanically ventilated, had higher positive-end-expiratory pressures and lower diastolic blood pressures and mean arterial pressures (p<0.015; Table 2). Most clinical signs reflecting organ perfusion differed between the groups: patients who died had significantly lower urine outputs, colder extremities, longer capillary refill times (CRT), and more severe skin mottling during the clinical examination.

Clinical examination and 90-day mortality

Both unadjusted and age- and sex- adjusted analyses of the overall population showed that most clinical examination findings were associated with 90-day mortality (Table S3). Multivariable logistic regression adjusted for age and noradrenaline infusion rate showed that five clinical examination findings, i.e. higher respiratory rate, higher systolic blood pressure, lower central temperature, altered consciousness, and decreased urine output were independently associated with 90-day mortality (Figure 2). The variables atrial fibrillation, diastolic blood pressure and severe skin mottling were of suggestive statistical significance because these variables had a p-value >0.015 and were statistically significant in less than 80 of the 100 bootstrap replications (Figure 2). The multivariable logistic regression analysis was repeated with systolic blood pressure in quartiles because this variable had a U-shaped relationship with mortality (Figure S1). In this model only the highest quartile (i.e. a systolic blood pressure > 133 mmHg) had a suggestive statistically significant association with mortality with an OR of 1.65 (98.5% CI 0.90-3.05; p=0.046; Table s4). In the complete case analysis (i.e., without imputed data) all associations except diastolic blood pressure remained statistically significant (Table S5).

(10)

153 Table 2. Clinical examination findings of survivors and non-survivors

Clinical Examination findings Survivors 90-day

N = 777 90-day non-survivors N = 298 P-value Age [years] 60 ± 15 67 ± 12 <0.001 Sex, male 480 (62%) 194 (65%) 0.31 Mechanical ventilation 424 (55%) 207 (69%) <0.001 PEEP [cm H2O] 7 (5, 8) 8 (5, 10) <0.001 Norepinephrine 345 (44%) 183 (61%) <0.001 Central circulation

Respiratory rate [per minute] 18 ± 6 19 ± 6 0.007

Heart rate [beats per minute] 87 ± 21 90 ± 22 0.06

Atrial fibrillation 42 (5%) 36 (12%) <0.001 SBP[mmHg] 119 ± 24 117 ± 27 0.33 DBP [mmHg] 60 ± 11 58 ± 12 0.007 MAP [mmHg] 79 ± 14 77 ± 15 0.017 CVP [mmHg] 8 (4, 12) 11 (8, 14) <0.001 Cardiac murmurs 71 (9%) 27 (9%) 0.97 Crackles or crepitations 100 (13%) 49 (16%) 0.13 Organ perfusion Consciousness Alert 264 (34%) 66 (22%) <0.001 Reacting to voice 162 (21%) 40 (13%) Reacting to pain 57 (7%) 32 (11%) Unresponsive 294 (38%) 160 (54%) Urine output [ml∙kg-1_∙h-1_] _{0.62 (0.34, 1.22) 0.42 (0.18, 0.83) <0.001} Urine output [ml∙kg-1_∙6h-1_] _{0.69 (0.40, 1.27) 0.51 (0.24, 0.90) <0.001} Central temperature [°C] 37.0 ± 0.9 36.8 ± 1.0 0.002 ΔTc-p, dorsum foot [°C] 7.5 ± 3.1 7.9 ± 3.3 0.07 ΔTc-p, big toe [°C] 9.2 ± 3.6 9.5 ± 3.7 0.18

Cold extremities, subjective 273 (35%) 129 (43%) 0.013

Capillary refill time sternum [s] 3.0 (2.0, 3.0) 3.0 (2.0, 4.0) <0.001

Capillary refill time finger [s] 2.0 (2.0, 4.0) 3.0 (2.0, 5.0) <0.001

Capillary refill time knee [s] 3.0 (2.0, 4.0) 4.0 (3.0, 5.0) <0.001

Skin mottling scorea

Mild (0-1) 560 (72%) 170 (57%) <0.001

Moderate (2-3) 201 (26%) 111 (37%)

Severe (4-5) 16 (2%) 17 (6%)

a_{Mottling was scored according to Ait-Oufella et al. (19). Abbreviations: OR, odds ratio; PEEP, positive end expiratory}

pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; CRT, capillary refill time; ΔTc-p, central-to-peripheral temperature difference.

Table 2. Clinical examination findings of survivors and non-survivors

a_{Mottling was scored according to Ait-Oufella et al.}19_{Abbreviations: OR, odds ratio; PEEP, positive end expiratory} pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; CRT, capillary refill time; ΔTc-p, central-to-peripheral temperature difference.

(11)

154

Performance of clinical examination when compared to prognostic scores

We compared our clinical examination model to three established ICU prognostic risk scores. When comparing AUC’s, the clinical examination model was comparable to the SAPS-II and APACHE-IV and significantly better than the SOFA score (Figure 3). The clinical examination model distinguished 817 (76%) patients correctly into survivor or non-survivor. The number of patients correctly classified was 810 (75%) for the SAPS-II, 818 (76%) for the APACHE-IV, 800 (74%) for the SOFA. The net reclassification improvement of the clinical examination model was 3.8% compared to the SAPS-II (p=0.09), 5.0% compared to the APACHE-IV (p=0.025), and 12% compared to the SOFA (p<0.001).

Sensitivity analysis: clinical examination and short-term mortality

The relation of clinical examination findings over time was studied using logistic regression analyses on 7- and 30-day mortality and a Cox regression. Severe skin mottling had stronger associations with 7-day mortality (OR 3.06; 95% CI 1.34 – 6.98; p=0.008) compared to 90-day mortality (OR 2.45; 95% CI 1.12 – 5.34; p=0.025). Systolic and diastolic blood pressures were not statistically significantly associated with 7-day mortality (Table S7). Atrial fibrillation and diastolic blood pressures were not statistically significantly associated with 30-day mortality (Table S7). Results of the multivariable Cox regression were comparable to the logistic regression analysis used in the main analysis (Table S8).

Subgroup analyses

In two predefined subgroup analyses the patient population was stratified by vasopressor use and by underlying pathology. In these analyses, only the eight clinical examination findings that were statistically significant in the primary analysis were tested (Figure S5). In patients without vasopressors, only higher respiratory rate and an altered consciousness had statistically significant associations with mortality (p<0.001; Table S9). In patients receiving vasopressors, higher respiratory rate, atrial fibrillation, lower central temperature, non-responsiveness, decreased urine output and severe skin mottling, were independently associated with 90-day mortality (Table S9). In patients admitted with septic shock only age and skin mottling over the knee were significantly associated with mortality (OR 3.22; 95% CI 1.31 - 7.94; p=0.011). In the subgroups of patients admitted with acute liver failure or post OLT, heart failure, with a CNS pathology, or after cardiac arrest, there were too few events (i.e. < 40) to assess any meaningful independent associations.

(12)

155 1. 21 (1. 12 - 1. 30) 0. 001 33 100 1. 13 (1. 02 - 1. 24) 0. 001 11 93 1. 30 (1. 10 - 1. 53) 0. 001 6. 5 99 1. 89 (0. 98 - 3. 64) 0. 017 5. 7 60 1. 11 (1. 01 - 1. 22) 0. 005 2. 2 83 1. 09 (0. 99 - 1. 20) 0. 001 3. 2 60 1. 25 (1. 02 1. 52) 0. 001 4. 8 80 Ref erence 16 0. 95 (0. 53 - 1. 71) 0.8 3 N /A 2. 73 (1. 36 - 5. 48) 0. 001 92 2. 39 (1. 47 - 3. 88) 0. 001 99 1. 44 (1. 09 - 1. 91) 0. 001 11 95 Ref erence 7.7 1. 29 (0. 86 - 1. 95) 0. 16 N /A 2. 52 (0. 94 - 6. 79) 0. 024 72

Age per 5 years

Norepinephrine dose per 0.

1 µg ⋅kg -1⋅h -1 Respirat ory rat

e per 5 min increase

At rial f ibrillat ion SB P per 10 mmHg increase DB P per 5 mmHg decrease Cent ral t emperat ure per °C decrease alert react ing t o voice react ing t o pain unresponsive Urine out put per ml ⋅kg -1⋅h -1 decrease mild moderat e severe Co vari ates Cen tral ci rcu lati on Or ga n pe rf us ion Consciousness Skin mot tling severit y 0 1 2 3 4 5 6 7 O dds rat io 98. 5% CI 95% CI OR (98.5% CI) P-value Independent contribution % Significant in # replications Figur e 2. Clinic al ex

amination findings associat

ed with 90-day mor talit y. F igur e description. F iv

e of the eight clinic

al ex

amination findings in our model w

er e independently associat ed with mor talit y (i.e . p < 0.015): r espir at or y r at e, syst olic blood pr essur e, c entr al t emper atur e, c onsciousness

, and urine output. T

he model

included all 1075 patients of whom 298 died

. P seudo R 2=0.14. Hosmer -L emesho w goodness -of-fit t est χ

2 7.01; p=0.72 (see plot in supplements

, e -F ig .4). A UC = 0.74 (95% CI 0.71-0.78). Mottling w as sc or ed ac cor ding t o A it-O ufella et al . 19 Abbr eviations: SBP , syst olic blood pr essur e; DBP , diast olic blood pr essur e.

(13)

156

Discussion

Clinical examination in 1075 adult patients acutely admitted to the ICU had reasonable prognostic accuracy. Five of the 19 tested clinical examination findings were independently associated with 90-day mortality. An increased respiratory rate, increased systolic blood pressure, lower core temperature, altered consciousness and decreased urine output had a statistically significant association with mortality. The predictive and discriminative value of a simple clinical examination approached that of the SAPS-II and APACHE-IV and outperformed the SOFA score.

In line with previous studies we found that clinical signs reflecting cerebral, renal, and skin hypoperfusion were independently associated with increased 90-day mortality in the critically ill.5,6,31 In our data severe skin mottling had a suggestively significant association with an OR of 2.48 (95% CI 1.13 –5.44), whereas others who assessed the persistence of skin mottling over time found a stronger association with an OR of 16 (95% CI 11 – 1568)19_{, and an OR of 3.29 (95% CI 2.08–} 5.19)4_{. The independent association of decreased urine output with 90-day mortality confirms} findings from the FINNAKI studies.5,32_{Similar to the modified early warning score (MEWS), we} found that an altered consciousness regardless of sedation significantly predicted mortality in the critically ill.33_{. All abovementioned variables may reflect the severity of critical illness on the} first day of ICU admission and as such identify patients at higher risk for circulatory failure and mortality. An alternative explanation might be that patients with these clinical signs are treated by interventions which worsen outcome.

The reasonable performance of our prediction models on 90-day mortality is in line with previous models derived from similar cohorts.34_{All prognostic scores performed worse than expected} from previous literature.14,15_{The inclusion criteria of the SICS-I may explain this discrepancy: we} studied 90-day mortality in patients acutely admitted to the ICU, whereas most prognostic scores perform best in evaluating in-hospital mortality or specific populations such as patients with trauma or suspected infection.35,36_{The use of an unselected population may produce unbiased} risk estimates and increases external validity.24_{The main disadvantage of this approach is that} average associations may be neutral or balanced out by different characteristics in different subgroups. The secondary analyses in clinically different subgroups were conducted to explore such associations: for example, a high systolic blood pressure was no longer independently associated with mortality in patients requiring vasopressors or in patients with septic shock (Table S9).

(14)

157 Figure 3. The discriminative value of the multivariable models to distinguish 90-day survivors from non- survivors using area under the receiver operating curves (AUC) analyses. Figure legend presents the area under the receiver-operating-curves with 98.5% confidence interval. The Delong method was used to compare our clinical examination model to three prognostic scores commonly used in the intensive care. Abbreviations: exam, examination; SAPS, simplified acute physiology score; APACHE, acute physiology and chronic health evaluation; SOFA, sequential organ failure assessment.

Implications and generalizability

The SICS-I provides evidence that a thorough clinical examination conducted on the first day of ICU admission may be used to obtain a rough estimation of 90-day mortality. By establishing the prognostic value of 19 clinical examination findings, we set the first step for a parsimonious clinical examination; i.e. the fewest number of clinical signs that yield the most prognostic value.35 These simple and easily obtainable clinical variables may better inform physicians in their clinical decision making. The examinations were conducted within 24 hours of ICU admission, usually in the morning, and after primary resuscitation efforts. There was no prespecified moment in time for the examination, which may decrease generalizability of the results. Nonetheless, this research practice does reflect daily clinical care where most patients are routinely assessed in the morning, regardless of the time that has passed since ICU admission.

Clinical examination for predicting mortality

0.00 0.25 0.50 0.75 1.00 Sensitivity 0.00 0.25 0.50 0.75 1.00 1−Specificity Clinical examination: 0.74 (0.70−0.78) SAPS−II: 0.76 (0.72−0.80; p=0.29) APACHE−IV: 0.77 (0.73−0.81; p=0.16) SOFA: 0.67 (0.63−0.72; p<0.001)

(15)

158

The dynamic care process of the critically ill patient may limit an accurate prediction of 90-day mortality with a single clinical examination or a prognostic score, which reflect a baseline mortality risk based on medical history and findings from the first 24 hours of ICU admission. The clinical status and treatment of a critically ill patient changes frequently and repeated clinical examinations might predict the individual patient prognosis more accurately. Previous studies have already shown that prolonged mechanical ventilation with high pressures or persistently low blood pressures, skin mottling, a decrease in urine output and increasing central-to-peripheral temperature gradients have strong associations with mortality.4,5,9,38-40_{Our clinical examination} was limited to a single time point, which could explain why not all these common prognostic variables were also statistically significant in the SICS-I. Future research should study the variation of clinical examination and associated interventions over time to assess its prognostic value.41 Strengths and limitations

The SICS-I was an unselected, single centre cohort of consecutive ICU patients and its findings require external validation in an independent cohort. To address this limitation, we assessed the robustness of our findings by adjusting for multiple outcomes, conducting multiple imputations and sensitivity analyses, and internally validating each predictive variable by bootstrap sampling. We evaluated a heterogeneous ICU population and certain prognostic associations may be more pronounced in patient subgroups, which is why we studied clinically relevant subgroups in our secondary analyses. Our findings do not apply to paediatric or electively admitted patients. The clinical examination findings collected in our study were not shared with caregivers. However, some of these findings (i.e. blood pressure, heart rate) were also assessed by caregivers and may have informed subsequent treatment decisions or were influenced by their interventions. The predictive value of a clinical variable measured at baseline therefore included the value of this variable combined with the subsequent intervention(s) to correct such a value. Since treatment strategies between physicians and countries differ, this fact may explain why different studies identify different predictors of mortality, in addition to population differences and other confounders. The prognostic value of clinical variables will be more transparent in a randomised setting where interventions are given based on different clinical treatment targets.37,38

Conclusion

A simple clinical examination, which can be performed in any critically ill patient in any setting, has reasonable discriminative value for assessing 90-day mortality in a single center cohort of acutely admitted ICU patients. In this study a single, protocolised clinical examination had similar prognostic abilities compared to the SAPS-II and APACHE-IV and outperformed the SOFA score.

(16)

159

References

Sakr Y, Vincent JL, Ruokonen E, et al. Sepsis and organ system failure are major determinants of post- intensive care unit mortality. J Crit Care. 2008;23(4):475-483.

Vincent JL, Marshall JC, Namendys-Silva SA, et al. Assessment of the worldwide burden of critical illness: The intensive care over nations (ICON) audit. Lancet Respir Med. 2014;2(5):380-386.

Macedo E, Malhotra R, Bouchard J, Wynn SK, Mehta RL. Oliguria is an early predictor of higher mortality in critically ill patients. Kidney Int. 2011;80(7):760-767.

Coudroy R, Jamet A, Frat JP, et al. Incidence and impact of skin mottling over the knee and its duration on outcome in critically ill patients. Intensive Care Med. 2015;41(3):452-459.

Vaara ST, Parviainen I, Pettila V, et al. Association of oliguria with the development of acute kidney injury in the critically ill. Kidney Int. 2016;89(1):200-208.

Houwink AP, Rijkenberg S, Bosman RJ, van der Voort, P H. The association between lactate, mean arterial pressure, central venous oxygen saturation and peripheral temperature and mortality in severe sepsis: A retrospective cohort analysis. Crit Care. 2016;20:56.

Hiemstra B, Eck RJ, Keus F, van der Horst, I C C. Clinical examination for diagnosing circulatory shock. Curr Opin Crit Care. 2017;23(4):293-301.

Cecconi M, De Backer D, Antonelli M, et al. Consensus on circulatory shock and hemodynamic monitoring. task force of the european society of intensive care medicine. Intensive Care Med. 2014;40(12):1795-1815. Dunser MW, Takala J, Ulmer H, et al. Arterial blood pressure during early sepsis and outcome. Intensive Care Med. 2009;35(7):1225-1233.

Varpula M, Tallgren M, Saukkonen K, Voipio-Pulkki LM, Pettila V. Hemodynamic variables related to outcome in septic shock. Intensive Care Med. 2005;31(8):1066-1071.

Champion HR, Sacco WJ, Hannan DS, et al. Assessment of injury severity: The triage index. Crit Care Med. 1980;8(4):201-208.

Ait-Oufella H, Bige N, Boelle PY, et al. Capillary refill time exploration during septic shock. Intensive Care Med. 2014;40(7):958-964.

de Moura EB, Amorim FF, da Cruz Santana, A N, et al. Skin mottling score as a predictor of 28-day mortality in patients with septic shock. Intensive Care Med. 2016;42(3):479-480.

Le Gall JR, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II) based on a european/ north american multicenter study. JAMA. 1993;270(24):2957-2963.

Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute physiology and chronic health evaluation (APACHE) IV: Hospital mortality assessment for today’s critically ill patients. Crit Care Med. 2006;34(5):1297- 1310.

Schriger DL, Baraff L. Defining normal capillary refill: Variation with age, sex, and temperature. Ann Emerg Med. 1988;17(9):932-935.

Curley FJ, Smyrnios NA. Routine monitoring of critically ill patients. In: Irwin RS, Cerra FB, Rippe JM, eds. Intensive care medicine. New York: Lippincott Williams & Wilkins; 2003:250-270.

Lima A, Jansen TC, van Bommel J, Ince C, Bakker J. The prognostic value of the subjective assessment of peripheral perfusion in critically ill patients. Crit Care Med. 2009;37(3):934-938.

Ait-Oufella H, Lemoinne S, Boelle PY, et al. Mottling score predicts survival in septic shock. Intensive Care Med. 2011;37(5):801-807.

Hiemstra B, Nolte IM, Tio CH, et al. Detailed statistical analysis plan of the simple intensive care studies-I (SICS-I). Clinicaltrials.gov: NCT02912624.

Little RJA. A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association. 1988;83(404):1198-1202.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

(17)

160

White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Stat Med. 2011;30(4):377-399.

Jakobsen JC, Gluud C, Wetterslev J, Winkel P. When and how should multiple imputation be used for handling missing data in randomised clinical trials - a practical guide with flowcharts. BMC Med Res Methodol. 2017;17(1):162.

Labarère J, Renaud B, Bertrand R, et al: How to derive and validate clinical prediction models for use in intensive care medicine. Intensive Care Med 2014;40:513-527

Afessa B, Gajic O, Keegan MT. Severity of illness and organ failure assessment in adult intensive care units. Crit Care Clin. 2007;23(3):639-658.

Razia Azen, Nicole Traxel. Using dominance analysis to determine predictor importance in logistic regression. Journal of Educational and Behavioral Statistics. 2009;34(3):319-347.

DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics. 1988;44(3):837-845.

Pencina MJ, D’Agostino RB S, D’Agostino RB, Vasan RS. Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond. Stat Med. 2008;27(2):12. Hiemstra B, Eck RJ, Koster G, et al. Clinical examination, critical care ultrasonography and outcomes in the critically ill: Cohort profile of the simple intensive care studies-I. BMJ Open. 2017;7(9):e017170.

Jakobsen JC, Wetterslev J, Winkel P, Lange T, Gluud C. Thresholds for statistical and clinical significance in systematic reviews with meta-analytic methods. BMC Med Res Methodol. 2014;14:120.

Champion HR, Sacco WJ, Hannan DS, et al. Assessment of injury severity: The triage index. Crit Care Med. 1980;8(4):201-208.

Varis E, Pettila V, Poukkanen M, et al. Evolution of blood lactate and 90-day mortality in septic shock. A post hoc analysis of the finnaki study. Shock. 2016;47(5):574-581.

Subbe CP, Kruger M, Rutherford P, Gemmel L. Validation of a modified early warning score in medical admissions. QJM. 2001;94(10):521-526.

Granholm A, Perner A, Krag M, et al. Development and internal validation of the simplified mortality score for the intensive care unit (SMS-ICU). Acta Anaesthesiol Scand. 2018;62(3):336-346.

Boyd CR, Tolson MA, Copes WS. Evaluating trauma care: The TRISS method. trauma score and the injury severity score. J Trauma. 1987;27(4):370-378.

Redfern OC, Smith GB, Prytherch DR, Meredith P, Inada-Kim M, Schmidt PE. A comparison of the quick sequential (sepsis-related) organ failure assessment score and the national early warning score in non-ICU patients with/without infection. Crit Care Med. 2018;46(12):1923-1933.

Sackett DL. The rational clinical examination. A primer on the precision and accuracy of the clinical examination. JAMA. 1992;267(19):2638-2644.

Damuth E, Mitchell JA, Bartock JL, Roberts BW, Trzeciak S. Long-term survival of critically ill patients treated with prolonged mechanical ventilation: A systematic review and meta-analysis. Lancet Respir Med. 2015;3(7):544-553. Accessed Mar 24, 2019.

Su L, Zhang R, Zhang Q, et al. The effect of mechanical ventilation on peripheral perfusion index and its association with the prognosis of critically ill patients. Crit Care Med. 2019. Accessed Mar 24, 2019. van Genderen ME, Bartels SA, Lima A, et al. Peripheral perfusion index as an early predictor for central hypovolemia in awake healthy volunteers. Anesth Analg. 2013;116(2):351-356.

Wiersema R, Castela Forte JN, Kaufmann T, et al. Observational study protocol for repeated clinical examination and critical care ultrasonography within the simple intensive care studies. J Vis Exp. 2019;16(143) 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

(18)

161 Pettila V, Merz T, Wilkman E, et al. Targeted tissue perfusion versus macrocirculation-guided standard care in patients with septic shock (TARTARE-2S): Study protocol and statistical analysis plan for a randomized controlled trial. Trials. 2016;17:x.

Hjortrup PB, Haase N, Bundgaard H, et al. Restricting volumes of resuscitation fluid in adults with septic shock after initial management: The CLASSIC randomised, parallel-group, multicentre feasibility trial. Intensive Care Med. 2016;42(11):1695-1705.

42

43

(19)

162

Su

pp

le

m

en

ta

ry

ta

bl

es

a

nd

fi

gu

re

s

Pr ev io us st ud ie s Ta bl e S1 . C lin ic al ex am in at io n fin di ng s as so ci at ed w ith 9 0‐ da y m or ta lit y id en tified in pr evious studies Au th or , y ea r N Populatio n Follow‐ up Adj uste d for Cl in ica l e xa m in at io n fin di ng s Age Respiratory rate/expans Blood pressure CVP Heart rate/rhythm Oliguria Mottling Capillary refill time Skin temperatur Altered sensorium Forrester, 1977 200 AMI Hospital discharge ‐ Champion, 1980 360 Criti cally injure d patients Hospital di sc har ge ‐ Pa rk er , 1 98 7 48 Se pt ic sh oc k IC U di sc ha rg e ‐ Tuchsmidt, 1989 145 Septic shock Hospital di sc har ge ‐ Bernardin, 1996 32 Septic shock Hospital discharge ‐ Hasdai, 1999 2478 Cardiogenic shoc k 30 days ‐ Varpula, 2005 111 Sepsis 30 days ‐ Dünser, 2009 274 Se ps is 28 days SA PS II Macedo, 2011 317 Critically ill ICU discharge Age, sepsis Ait‐Oufe lla, 201 1 60 Septic shock 14 days ‐ Van Gend eren, 2012 25 OHCA 28 days ‐ De Backer, 2013 252 Severe sepsis ICU discharge APACHE II, SOFA Coudroy, 2015 791 Critically ill ICU discharge SAPS II Ait‐Oufe lla, 201 4 59 Septic shock 14 days ‐ De Moura, 2016 97 Septic shock 28 days Age, SOFA Vaara, 2016 1966 Cr iti ca lly il l 90 days Age, APACHE II, SAPS II Houwink, 2016 821 Critically ill with infe ction Hospital dischar ge Ag e, APACHE IV Varis, 2016 496 Septic shock 90 days SAPS II, SOFA Table legend . = significant pre dictor in univariab le analysis. = sig nificant pre dicto r in mul tivariabl e analysis. = ver y si gnificant (p<0.015) p redictor in multivariable analysis. = nonsignificant (p>0 .05) predictor in univariable a nalysis. = nonsignificant predi ctor in multivariable analysis. Abbreviat ions: AMI, acute my ocardial infa rction; OHCA, ou t‐o f‐hos pi ta l c ar di ac a rr es t; IC U, intensive care unit. Supplemen tar y app

endix: additional tables and figur

es Pr evious studies Table S1. Clinic al ex

amination findings associat

ed with 90-day mor talit y identified in pr evious studies Table legend . = signific ant pr edict or in univ ariable analysis . = signific ant pr edict or in m ultiv ariable analysis . = v er y signific ant (p<0.015) pr edict or in m ultiv ariable analysis . = nonsignific ant (p>0.05) pr edict or in univ ariable analysis . = nonsignific ant pr edict or in m ultiv ariable analysis . Abbr

eviations: AMI, acut

e my oc ar dial inf ar

ction; OHCA,

out-of-hospital c

ar

diac arr

est; ICU, int

ensiv

e c

ar

(20)

163 Descriptives and missing percentages

Figure S1. Graphical representation of bootstrap sampling

Explanation: for each bootstrap sample, 1075 patients were drawn one by one from the imputed dataset with replacement. Drawing with replacement means that each selected case was placed back into the dataset so that it was available for the next draw. These 100 bootstrap samples are comparable, but not identical to the original data set. This process mimics the sampling from an underlying population.

(21)

164

Descriptives and missing percentage

Table S2. Description of collected variables and percentage of missingness

Variable Method of measuring Missing %

Age, years Patient’s electronic chart 0.0

Sex, male Patient’s electronic chart 0.0

Mechanical ventilation

Inspiratory O2, % mechanical ventilator 0.0

PEEP, cm H2O mechanical ventilator 0.0

Vasoactive medication, µg∙kg‐1_∙min‐1 _{Infusion pump inspection at bedside} _0.0

Clinical examination Central circulation Respiratory rate, per minute bedside (electrocardiographic) monitor 0.0 Heart rate, beats per minute bedside (electrocardiographic) monitor 0.0 Heart rhythm bedside (electrocardiographic) monitor 0.0 Systolic blood pressure, mmHg arterial line and sphygmomanometer 0.2 Diastolic blood pressure, mmHg arterial line and sphygmomanometer 0.2 Mean arterial pressure, mmHg arterial line and sphygmomanometer 0.3 Central venous pressure, mmHg central venous line in internal jugular, subclavian or femoral vein 76.6* Cardiac murmurs Auscultation at the 2nd_{intercostal space left and}

right, 4th_or 5th_{intercostal space left and apex}

9.6 Pulmonary crepitations or crackles Auscultation of the chest at the superior, inferior and basal lung fields 1.3 Tissue perfusion Mental state (alert/voice/pain/ unresponsive) Observation 0.0

Sedative medication, µg∙kg‐1_∙min‐1 _{Infusion pump inspection at bedside} _4.9

Urine output, ml∙kg‐1_∙h‐1 _{Patient’s electronic chart} _1.9 Central temperature, °C Bladder thermistor catheter 1.5 Skin temperature dorsum foot, °C skin temperature of dorsum foot using skin probe 15.3 Skin temperature big toe, °C skin temperature of big toe using skin probe 9.9 Δ‐Temperature Bladder to dorsum foot or big toe difference 15.3 Cold extremities, cold or warm Subjective assessment 0.6 Capillary refill time sternum, s Palpation 13.7 Capillary refill time finger, s Palpation 2.5 Capillary refill time knee, s Palpation 9.9 Mottling score, 0 to 5 Observation of mottling area on the legs, scoring system according to Ait‐Oufella19 10.1 *Central venous pressure was missing in > 50% of the patients and was excluded from the multiple imputations.

Table S2. Description of collected variables and percentage of missingness

(22)

165 Primary analysis

Table S3. Associations of clinical examination findings on 90-day mortality

Left: crude OR’s. Right: OR’s adjusted for age and sex. Abbreviations: OR, odds ratio; PEEP, positive end expiratory pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; CRT, capillary refill time; ΔTc-p, central-to-peripheral temperature difference.

4 eTable 3. Associations of clinical examination findings on 90-day mortality

Variable OR 95% CI P-value OR 95% CI P-value

UNADJUSTED AGE- AND SEX- ADJUSTED

Age per year increase 1.04 1.03 - 1.06 <0.001 - - -

Gender male 1.15 0.87 - 1.53 0.313 - - -

Mechanical ventilation 1.89 1.43 - 2.52 <0.001 2.01 1.50 - 2.70 <0.001 PEEP per cm H2O increase 1.12 1.05 - 1.20 0.001 1.14 1.06 - 1.23 <0.001 Noradrenaline dose, µg∙kg-1_∙min-1 _5.85 _{2.95 - 11.6} _<0.001 _7.18 _{3.53 - 14.6} _<0.001

Clinical examination

Central circulation

Respiratory rate per minute increase 1.03 1.01 - 1.06 0.008 1.03 1.01 - 1.06 0.006 Heart rate per minute increase 1.01 1.00 - 1.01 0.060 1.01 1.00 - 1.01 0.012 Atrial fibrillation 2.40 1.51 - 3.84 <0.001 1.48 0.91 - 2.42 0.114 SBP per mmHg increase 1.00 0.99 - 1.00 0.352 1.00 0.99 - 1.00 0.205 DBP per mmHg decrease 1.02 1.00 - 1.03 0.009 1.01 1.00 - 1.02 0.065 MAP per mmHg decrease 1.01 1.00 - 1.02 0.019 1.01 1.00 - 1.02 0.029

Cardiac murmurs 1.01 0.63 - 1.61 0.970 0.93 0.57 - 1.51 0.779

Crackles or crepitations 1.34 0.92 - 1.94 0.122 1.22 0.83 - 1.78 0.315 Organ perfusion

Consciousness

alert 1.00 Reference - 1.00 Reference

reacting to voice 0.99 0.64 - 1.53 0.956 0.95 0.60 - 1.49 0.819 reacting to pain 2.25 1.35 - 3.74 0.002 2.34 1.38 - 3.99 0.002 unresponsive 2.18 1.56 - 3.03 <0.001 2.33 1.65 - 3.28 <0.001 Urine output per ml∙kg-1_∙h-1_decrease _1.47 _{1.21 - 1.77} _<0.001 _1.41 _{1.16 - 1.70} _<0.001 Urine output per ml∙kg-1_∙6h-1_{decrease 1.66} _{1.33 - 2.07} _<0.001 _1.52 _{1.22 - 1.90} _<0.001 Central temperature per °C decrease 1.26 1.09 - 1.46 0.002 1.23 1.05 - 1.43 0.010 ΔTc-p, dorsum foot per °C increase 1.04 0.99 - 1.08 0.087 1.00 0.97 - 1.06 0.612 ΔTc-p, toe per °C increase 1.02 0.99 - 1.06 0.208 1.00 0.96 - 1.04 0.926 Cold extremities, subjective 1.41 0.54 - 0.93 0.015 1.22 0.62 - 1.09 0.171 CRT sternum per second increase 1.29 1.15 - 1.46 <0.001 1.24 1.09 - 1.40 0.001 CRT finger per second increase 1.17 1.10 - 1.25 <0.001 1.14 1.07 - 1.21 <0.001 CRT knee per second increase 1.17 1.11 - 1.24 <0.001 1.14 1.08 - 1.21 <0.001 Mottling rate

Mild (0-1) 1.00 Reference - 1.00 Reference

Moderate (2-3) 1.75 1.29 - 2.38 <0.001 1.58 1.16 - 2.16 0.004 Severe (4-5) 3.63 1.66 - 6.75 0.001 3.42 1.65 - 7.07 0.001 Left: crude OR’s. Right: OR’s adjusted for age and sex. Abbreviations: OR, odds ratio; PEEP, positive end expiratory pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; CRT, capillary refill time; ΔTc-p, central-to-peripheral temperature difference.

(23)

166 0 .2 .4 .6 .8 Actual probability 0 .2 .4 .6 Predicted probability

observed (proportion) predicted (proportion)

.2 .3 .4 .5 .6

Mortality at 90−day follow−up

50 100 150 200

Systolic blood pressure (mmHg)

Figure S3. U-shaped association of systolic blood pressure with mortality using penalized spline regression

Figure S4. Calibration plot of the clinical examination: observed versus predicted mortality across 10 equally sized groups

Comment: the actual probability deviates little from the predicted probability. Combined with a non-significant Hosmer Lemeshow test (χ2_{7.01; p=0.72), we judged these deviations as acceptable.}

(24)

167 Table S4. Clinical examination model with systolic blood pressure in quartiles

The model included all 1075 patients of whom 298 died. Pseudo R2=0.13. Hosmer-Lemeshow goodness-of-fit test χ2 7.01; p=0.72. AUC = 0.74 (95% CI 0.71-0.78). Abbreviations: OR, odds ratio; CI, confidence interval.

Table S5. Clinical examination model on complete cases (i.e. only patients without missing data)

The model included all 916 patients of whom 257 died. Pseudo R2=0.14. Hosmer-Lemeshow goodness-of-fit test χ2 9.02; p=0.52. AUC = 0.75 (95% CI 0.71-0.78). Abbreviations: OR, odds ratio; CI, confidence interval.

5

eTable 4. Clinical examination model with systolic blood pressure in quartiles

Variable OR 95% CI P-value

Age in years 1.04 1.03 - 1.05 <0.001

Noradrenaline per µg∙kg-1_∙min-1 _3.15 _{1.45 - 6.85} _0.004

Respiratory rate per minute 1.05 1.02 - 1.08 <0.001

Atrial fibrillation 1.78 1.05 - 3.00 0.031

Systolic blood pressure per quartile increase

1st_{quartile (< 100 mmHg) 1.00} _Reference

2nd_{quartile (100 - 114 mmHg) 1.17} _{0.77 - 1.78} _0.452

3rd_{quartile (114 - 133 mmHg) 1.04} _{0.68 - 1.61} _0.846

4th_{quartile (> 133 mmHg) 1.65} _{1.01 - 2.71} _0.046

Diastolic blood pressure per mmHg decrease 1.01 1.00 - 1.03 0.072

Central temperature per °C decrease 1.22 1.04 - 1.44 0.012

Organ perfusion Consciousness Alert 0.93 0.58 - 1.50 0.779 Reacting to voice 2.75 1.57 - 4.81 <0.001 Reacting to pain 2.30 1.55 - 3.39 <0.001 Unresponsive 1.43 1.14 - 1.78 0.002

Urine output per µg∙kg-1_∙6h-1 _decrease

Skin mottling severity 1.00 Reference

Mild (0-1) 1.27 0.91 - 1.77 0.157

Moderate (2-3) 2.48 1.12 - 5.50 0.026

Severe (4-5) 1.04 1.03 - 1.05 <0.001

The model included all 1075 patients of whom 298 died. Pseudo R2_{=0.13. Hosmer-Lemeshow goodness-of-fit test χ}2

7.01; p=0.72. AUC = 0.74 (95% CI 0.71-0.78). Abbreviations: OR, odds ratio; CI, confidence interval.

6

eTable 5. Clinical examination model on complete cases (i.e. only patients without missing data)

Clinical examination OR 95% CI

P-value

Age in years 1.04 1.03 - 1.06 <0.001

Noradrenaline per µg∙kg-1_∙min-1 _4.51 _{1.88 - 10.8} _0.001

Systolic blood pressure per mmHg increase 1.01 1.00 - 1.02 0.009

Central temperature per °C decrease 1.00 1.00 - 1.00 <0.001

Organ perfusion Consciousness Alert 1.00 Reference Reacting to voice 3.07 1.70 - 5.57 <0.001 Reacting to pain 2.36 1.53 - 3.62 <0.001 Unresponsive 1.33 1.06 - 1.68 0.016

Urine output per µg∙kg-1_∙h-1 _decrease _1.24 _{1.05 - 1.47} _0.013

Skin mottling severity

Mild (0-1) 1.00 Reference

Moderate (2-3) 1.32 0.94 - 1.86 0.112

Severe (4-5) 2.90 1.27 - 6.62 0.012

The model included all 916 patients of whom 257 died. Pseudo R2_{=0.14. Hosmer-Lemeshow goodness-of-fit test χ}2

(25)

168

Sensitivity analysis

Table S6. Associations of clinical examination on 7- and 30-day mortality.

Abbreviations: OR, odds ratio; PEEP, positive end expiratory pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; ΔTc-p, central-to-peripheral temperature difference; CRT, capillary refill time. 11 Sensitivity analysis Table S6. Associations of clinical examination on 7‐ and 30‐day mortality. Variable OR (95% CI) 7‐days P‐value OR (95% CI) 30‐days P‐value Age, year increase 1.04 (1.03 ‐ 1.06) <0.001 1.04 (1.03 ‐ 1.05) <0.001 Sex, male 1.08 (0.75 ‐ 1.55) 0.67 1.05 (0.78 ‐ 1.40) 0.75 Mechanical ventilation 1.73 (1.19 ‐ 2.50) 0.004 1.87 (1.39 ‐ 2.53) <0.001 PEEP, cm H2O increase 1.13 (1.05 ‐ 1.23) 0.002 1.15 (1.07 ‐ 1.23) <0.001 Noradrenaline dose, µg∙kg‐1_∙min‐1 _{4.76 (2.31 ‐ 9.82) <0.001 4.84 (2.47 ‐ 9.49)} _<0.001 Central circulation Respiratory rate, minute increase 1.04 (1.01 ‐ 1.07) 0.003 1.03 (1.01 ‐ 1.06) 0.009 Heart rate, minute increase 1.01 (1.00 ‐ 1.01) 0.11 1.00 (1.00 ‐ 1.01) 0.15 Atrial fibrillation 2.46 (1.44 ‐ 4.19) 0.001 2.28 (1.41 ‐ 3.67) 0.001 SBP, mmHg increase 0.99 (0.98 ‐ 1.00) 0.037 0.99 (0.99 ‐ 1.00) 0.08 DBP, mmHg decrease 1.02 (1.00 ‐ 1.03) 0.028 1.02 (1.00 ‐ 1.03) 0.009 MAP, mmHg decrease 1.02 (1.00 ‐ 1.03) 0.014 1.01 (1.00 ‐ 1.02) 0.018 Cardiac murmurs 1.03 (0.57 ‐ 1.87) 0.91 1.10 (0.68 ‐ 1.79) 0.69 Crackles or crepitations 0.76 (0.45 ‐ 1.31) 0.33 1.06 (0.71 ‐ 1.59) 0.77 Organ perfusion Consciousness

alert 1.00 (Reference) 1.00 (Reference) reacting to voice 0.98 (0.55 ‐ 1.73) 0.94 1.08 (0.68 ‐ 1.72) 0.73 reacting to pain 1.07 (0.51 ‐ 2.25) 0.87 1.85 (1.06 ‐ 3.20) 0.029 unresponsive 1.94 (1.27 ‐ 2.96) 0.002 2.23 (1.57 ‐ 3.17) <0.001 Urine output, ml∙kg‐1_∙h‐1_decrease _{1.62 (1.22 ‐ 2.15)} _0.001 _{1.54 (1.25 ‐ 1.91)} _<0.001 Urine output, ml∙kg‐1_∙6h‐1_decrease _{1.81 (1.32 ‐ 2.47)} _0.000 _{1.64 (1.29 ‐ 2.07)} _<0.001 Central temperature, °C decrease 1.28 (1.06 ‐ 1.55) 0.009 1.31 (1.12 ‐ 1.54) 0.001 ΔTc‐p, dorsum of foot, °C increase 1.06 (1.00 ‐ 1.12) 0.033 1.04 (0.99 ‐ 1.09) 0.11 ΔTc‐p, toe, °C increase 1.05 (1.00 ‐ 1.10) 0.044 1.02 (0.98 ‐ 1.06) 0.27 Cold extremities, subjective 1.58 (1.12 ‐ 2.25) 0.010 1.37 (1.03 ‐ 1.83) 0.030 CRT sternum, second increase 1.30 (1.13 ‐ 1.50) <0.001 1.22 (1.08 ‐ 1.38) 0.002 CRT finger, second increase 1.15 (1.07 ‐ 1.23) <0.001 1.12 (1.06 ‐ 1.19) <0.001 CRT knee, second increase 1.10 (1.04 ‐ 1.17) 0.001 1.11 (1.05 ‐ 1.17) <0.001 Mottling rate

Mild (0‐1) 1.00 (Reference) 1.00 (Reference) Moderate (2‐3) 1.69 (1.14 ‐ 2.54) 0.009 1.74 (1.26 ‐ 2.39) 0.001

Severe (4‐5) 4.41 (2.10 ‐ 9.34) <0.001 2.94 (1.44 ‐ 5.98) 0.003

Abbreviations: OR, odds ratio; PEEP, positive end expiratory pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; ΔTc‐p, central‐to‐peripheral temperature difference; CRT, capillary refill time.

(26)

169 Table S7. Clinical examination independently associated with 7-day (left) and 30-day (right)

The7-day mortality model included 1075 patients of whom 151 died. Pseudo R2_{=0.11. Hosmer-Lemeshow} goodness- of-fit test χ2_{4.83; p=0.90. AUC = 0.74 (95% CI 0.70-0.78). The 30-day mortality model included 1075} patients of whom 255 died. Pseudo R2_{=0.11. Hosmer-Lemeshow goodness-of-fit test χ}2_{8.07; p=0.62. AUC = 0.73} (95% CI 0.70-0.76). Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure.

Table S8. Cox regression of clinical examination findings independently associated with 90-day mortality

The model included all 1075 patients of whom 298 died. Mottling was scored according to Ait-Oufella et al.19 Clinical examination for predicting mortality

9

eTable 8. Cox regression of clinical examination findings independently associated with 90-day

mortality

Clinical examination finding Hazard

ratio 95% confidence interval P-value

Age in years 1.03 1.02 - 1.04 <0.001

Noradrenaline per µg∙kg-1_∙min-1 _2.21 _{1.31 - 3.74} _0.003

Systolic blood pressure per mmHg increase 1.01 1.00 - 1.01 0.037

Central temperature per °C decrease 1.20 1.06 - 1.36 0.005

Organ perfusion Consciousness Alert 1.00 Reference Reacting to voice 1.00 0.67 - 1.49 0.991 Reacting to pain 2.18 1.41 - 3.38 <0.001 Unresponsive 2.01 1.46 - 2.76 <0.001

Urine output per µg∙kg-1_∙h-1 _decrease _1.38 _{1.14 - 1.67} _0.001

Skin mottling severity

Mild (0-1) 1.00 Reference

Moderate (2-3) 1.27 0.98 - 1.64 0.067

Severe (4-5) 1.92 1.14 - 3.25 0.015

The model included all 1075 patients of whom 298 died. Mottling was scored according to Ait-Oufella et al.19

Table S7. Clinical examination independently associated with 7‐day (left) and 30‐day (right) mortality

Variable OR (95% CI) P‐value OR (95% CI) P‐value

Age in years 1.04 (1.02 ‐ 1.05) <0.001 1.03 (1.02 ‐ 1.04) <0.001 Noradrenaline per µg∙kg‐1_∙min‐1 _{2.76 (1.15 ‐ 6.62)} _0.022 _{2.64 (1.21 ‐ 5.74)} _0.014 Central circulation Respiratory rate per minute 1.06 (1.02 ‐ 1.09) 0.001 1.05 (1.02 ‐ 1.08) <0.001 Atrial fibrillation 1.89 (1.05 ‐ 3.41) 0.035 1.85 (1.09 ‐ 3.16) 0.023 SBP per mmHg increase 1.00 (0.99 ‐ 1.01) 0.51 1.01 (1.00 ‐ 1.01) 0.09 DBP per mmHg decrease 0.99 (0.97 ‐ 1.01) 0.37 1.02 (1.00 ‐ 1.03) 0.07 Central temperature per °C decrease 0.82 (0.68 ‐ 0.99) 0.044 1.28 (1.09 ‐ 1.51) 0.003 Organ perfusion Consciousness 1.00 (Reference) Alert 0.93 (0.51 ‐ 1.70) 0.81 1.00 (Reference) Reacting to voice 1.20 (0.54 ‐ 2.62) 0.66 1.08 (0.66 ‐ 1.76) 0.77 Reacting to pain 1.94 (1.19 ‐ 3.17) 0.008 2.19 (1.21 ‐ 3.95) 0.010 Unresponsive 1.45 (1.07 ‐ 1.97) 0.017 2.37 (1.58 ‐ 3.56) <0.001 Urine output per µg∙kg‐1_∙6h‐1 _decrease _{1.40 (1.11 ‐ 1.77)} _0.005 Skin mottling score

Mild (0‐1) 1.00 (Reference) 1.00 (Reference)

Moderate (2‐3) 1.25 (0.81 ‐ 1.91) 0.32 1.28 (0.91 ‐ 1.81) 0.15

Severe (4‐5) 3.06 (1.34 ‐ 6.98) 0.008 2.04 (0.93 ‐ 4.50) 0.07

The7‐day mortality model included 1075 patients of whom 151 died. Pseudo R2_{=0.11. Hosmer‐Lemeshow goodness‐}

of‐fit test χ2_{4.83; p=0.90. AUC = 0.74 (95% CI 0.70‐0.78). The 30‐day mortality model included 1075 patients of whom}

255 died. Pseudo R2_{=0.11. Hosmer‐Lemeshow goodness‐of‐fit test χ}2_{8.07; p=0.62. AUC = 0.73 (95% CI 0.70‐0.76).}

Figures S3 and S4 present information on model calibration and discrimination. Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure.